parthknowsai
Region: US
Wednesday 02 September 2026 12:24:24 GMT
49415
2838
76
226
Music
Download
Comments
fomalqaut :
fraction of data, tenfold more compute, great job 👍
2026-09-02 16:50:42
138
Silver Sunshine :
Nice discussion
2026-09-03 12:57:48
0
orangehipo :
How they measure that the recommendations “got better” ?
2026-09-02 18:53:07
40
Salmon :
Bro they have 5 shows, how hard can it be. Also, they cancel shows without ending it properly, that’s bit I really fokin hate.
2026-09-03 07:26:54
0
Inspector Mango Bango :
i bet i single handily can create algorithm like that
2026-09-03 07:58:24
0
Darklove :
so this is why I can not find anything good on Netflix
2026-09-03 08:59:25
1
ClashofAsh :
I swear that old system just recommended everything that was new or currently going on. Hopefully, this new system will actually work.
2026-09-02 15:43:51
66
tik_tok_shau :
like google, any llm is better than they search tool
2026-09-03 08:46:11
0
️Abdelrahman Elattar :
China made is simple, cheap and efficient.🤝
2026-09-02 15:19:09
4
BusyDoggy 🇪🇺 :
I mean... it's not hard to improve Netflix's recommendations 😂 I have wondered for years what all those people do there because it's so terrible.
2026-09-03 10:55:41
2
Loh :
It shows us that LLM have uses but not able to do everything so it might not be dead but it doesn’t have as much power as we initially thought 🤷🏽♂️
2026-09-02 12:38:21
50
Gabriel Kemmer :
Now makes sense why I’ve uninstalled Netflix
2026-09-03 10:50:15
0
Yene :
What do you mean it does not obviously scale? They were doing fine until LLM came about.
2026-09-03 05:26:03
1
J :
Netflix has never recommended anything worth watching to me.
2026-09-03 07:27:27
2
@krakrak🏳️🌈🇵🇱🇪🇺 :
Why would they use thousands of hand made features instead of using standard ai recomendation algorithms?
2026-09-03 01:16:41
0
Kike :
I tried the Netflix ai and it doesnt work
2026-09-03 08:38:28
1
Mathilde :
I always wondered about the strange recommendations of Netflix. It rarely fitted my interests. So hopefully roll the LLM based version out
2026-09-03 03:46:33
6
Gogomeow :
but still recommendation doesn't work for me
2026-09-02 21:57:52
2
666bbb.0 :
Watch Netflix. Buy from Amazon. Order from DoorDash. Never leave the house. Start hoarding. 😂😂
2026-09-02 15:40:27
1
bagienny :
In my opinion, netflix’s algorithms were just marketing hype. There was no basis in reality; the only differences were in the details, I was shown more blue images, whilst my friends were shown different shades, and sometimes the order differed. Regardless of gender, age, preferences or ratings we’d given, we were all shown the same suggestions. It’s much the same with LLMs: before you get to the nitty-gritty or any actual solutions, you have to wade through the statistics, which are the very essence of LLMs, because what they recommend is like google search, it shows you what it wants to show you, not what you actually want.
2026-09-03 06:40:27
2
Speech :
not that an LLM is better than semantic search, it’s just that the old system sucked!
2026-09-02 17:52:05
3
Matt 🇫🇮 :
💯 is that working on Netflix now?
2026-09-02 12:31:49
0
Perunov :
They also removed old star system with simplified yes/no 🤷♂️and yes, of course pattern recognition system would be able to recognize similarities between content 😂
2026-09-02 15:30:52
2
PanoςΕ :
all these but they didn't build the algorithm for: watched list or continue watching list. 🤡
2026-09-03 02:34:20
1
Nateu :
will it finally stop recommending things I’ve already watched? And go and grab the things I loved 10+ years ago and actually want to rewatch?
2026-09-02 19:11:31
1
To see more videos from user @parthknowsai, please go to the Tikwm
homepage.
